Annotation of parser3/src/include/pa_string.h, revision 1.145.4.9
1.41 paf 1: /** @file
1.43 paf 2: Parser: string class decl.
3:
1.145 paf 4: Copyright (c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com)
1.124 paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.1 paf 6: */
7:
8: #ifndef PA_STRING_H
9: #define PA_STRING_H
1.140 paf 10:
1.145.4.9! paf 11: static const char* IDENT_STRING_H="$Date: 2003/09/24 11:52:05 $";
1.145 paf 12:
13: // includes
1.1 paf 14:
1.4 paf 15: #include "pa_types.h"
1.145 paf 16: #include "pa_array.h"
17:
18: extern "C" { // cord's author forgot to do that
19: #define CORD_NO_IO
20: #include "cord.h"
21: };
1.4 paf 22:
1.145.4.1 paf 23: // cord extension
24: /* Returns true if x does contain */
25: /* char not_c at positions i..i+n. Value i,i+n must be < CORD_len(x). */
1.145.4.5 paf 26: int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c);
1.145.4.1 paf 27:
1.145 paf 28: // forwards
1.9 paf 29:
1.145 paf 30: class Charset;
1.135 paf 31: class Table;
1.71 paf 32: class SQL_Connection;
1.101 parser 33: class Dictionary;
1.145 paf 34: class Request_charsets;
35: class String;
36: typedef Array<const String*> ArrayString;
37:
38: /// this is result of pos functions which mean that substr were not found
39: #define STRING_NOT_FOUND ((size_t)-1)
40:
41: class StringBody {
42:
43: CORD body;
44:
45: public:
46:
47: StringBody(): body(CORD_EMPTY) {}
48: StringBody(CORD abody): body(abody) {
49: assert(!body // no body
50: || *body // ordinary string
51: || body[1]==1 // CONCAT_HDR
52: || body[1]==4 // FN_HDR
53: || body[1]==6 // SUBSTR_HDR
54: );
55: }
56: /// WARNING: length is only HELPER length, str in ANY case should be zero-terminated
57: StringBody(const char* str, size_t helper_length): body(CORD_EMPTY) {
58: append_know_length(str, helper_length?helper_length:strlen(str));
59: }
60: static StringBody Format(int value);
61:
62: void clear() { body=CORD_EMPTY; }
63:
64: bool operator! () const { return is_empty(); }
65:
66: uint hash_code() const;
67:
68: const char* cstr() const { return CORD_to_const_char_star(body); }
69: char* cstrm() const { return CORD_to_char_star(body); }
70:
71: size_t length() const { return CORD_len(body); }
72:
73: bool is_empty() const { return body==CORD_EMPTY; }
74:
75: void append_know_length(const char *str, size_t known_length) {
76: if(known_length)
77: body=CORD_cat_char_star(body, str, known_length);
78: }
79: void append_strdup_know_length(const char* str, size_t known_length) {
80: if(known_length)
81: append_know_length(pa_strdup(str, known_length), known_length);
82: }
83: void append(char c) { body=CORD_cat_char(body, c); }
84: StringBody& operator << (const StringBody src) { body=CORD_cat(body, src.body); return *this; }
85: StringBody& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; }
86:
87: // could not figure out why this operator is needed [should do this chain: string->simple->==]
88: bool operator < (const StringBody src) const { return CORD_cmp(body, src.body)<0; }
89: bool operator > (const StringBody src) const { return CORD_cmp(body, src.body)>0; }
90: bool operator <= (const StringBody src) const { return CORD_cmp(body, src.body)<=0; }
91: bool operator >= (const StringBody src) const { return CORD_cmp(body, src.body)>=0; }
92: bool operator != (const StringBody src) const { return CORD_cmp(body, src.body)!=0; }
93: bool operator == (const StringBody src) const { return CORD_cmp(body, src.body)==0; }
94:
95: int ncmp(size_t x_begin, const StringBody y, size_t y_begin, size_t size) const {
96: return CORD_ncmp(body, x_begin, y.body, y_begin, size);
97: }
98:
99: char fetch(size_t index) const { return CORD_fetch(body, index); }
100: StringBody mid(size_t index, size_t length) const { return CORD_substr(body, index, length); }
101: size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); }
102: size_t pos(const StringBody substr, size_t offset=0) const {
103: if(!substr.length())
104: return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
105: return CORD_str(body, offset, substr.body);
106: }
107: size_t pos(char c,
108: size_t offset=0) const {
109: return CORD_chr(body, offset, c);
110: }
111:
1.145.4.1 paf 112: /* template<typename I> void for_each(int (*callback)(const char* s, I), I info) const {
1.145 paf 113: CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info);
1.145.4.1 paf 114: }*/
1.145 paf 115:
116: void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
117:
1.145.4.2 paf 118: /*StringBody normalize() const {
1.145 paf 119: return StringBody(CORD_balance(body));
1.145.4.2 paf 120: }*/
1.145 paf 121:
122: void dump() const {
123: CORD_dump(body);
124: }
125: };
1.62 paf 126:
1.145.4.4 paf 127: template<typename T>
128: inline size_t get_length(T current) {
129: return current;
130: }
1.42 paf 131: /**
1.145.4.1 paf 132: String which knows the lang of all it's langs.
1.41 paf 133:
134: All pieces remember
135: - whether they are tainted or not,
1.145.4.1 paf 136: and the lang which should be used to detaint them
1.41 paf 137: */
1.145 paf 138: class String: public PA_Object {
1.1 paf 139: public:
1.48 paf 140:
1.145.4.1 paf 141: /** piece is tainted or not. the lang to use when detaint
1.106 parser 142: remember to change String_Untaint_lang_name @ untaint.C along
143: */
1.145 paf 144: enum Language {
1.145.4.1 paf 145: L_UNSPECIFIED=0, ///< no real string has parts of this lange: it's just convinient to check when string's empty
1.145 paf 146: // these two must go before others, there are checks for >L_AS_IS
1.145.4.7 paf 147: L_CLEAN='1', ///< clean
1.145 paf 148: L_AS_IS, ///< leave all characters intact
1.122 paf 149:
1.145 paf 150: L_PASS_APPENDED,
1.41 paf 151: /**<
1.145.4.1 paf 152: leave lang built into string being appended.
1.41 paf 153: just a flag, that value not stored
154: */
1.145.4.1 paf 155: L_TAINTED, ///< tainted, untaint lang as assigned later
156: // untaint langs. assigned by ^untaint[lang]{...}
1.145 paf 157: L_FILE_SPEC, ///< file specification
158: L_HTTP_HEADER, ///< text in HTTP response header
159: L_MAIL_HEADER, ///< text in mail header
160: L_URI, ///< text in uri
161: L_TABLE, ///< ^table:set body
162: L_SQL, ///< ^table:sql body
163: L_JS, ///< JavaScript code
164: L_XML, ///< ^dom:set xml
165: L_HTML, ///< HTML code (for editing)
1.145.4.1 paf 166: L_OPTIMIZE_BIT = 0x80 ///< flag, requiring cstr whitespace optimization
1.27 paf 167: };
168:
1.145.4.1 paf 169: class Languages {
170:
171: union {
172: struct {
173: Language lang:8;
1.145.4.5 paf 174: int is_not_just_lang:sizeof(CORD)*8-8;
1.145.4.1 paf 175: };
176: CORD langs;
177: };
178:
1.145.4.4 paf 179: template<typename C>
180: CORD make_langs(C current) const {
1.145.4.1 paf 181: return is_not_just_lang?
182: langs
1.145.4.4 paf 183: :CORD_chars((char)lang, get_length(current));
1.145 paf 184: }
185:
1.145.4.1 paf 186: CORD make_langs(size_t aoffset, size_t alength) const {
187: return is_not_just_lang?
188: CORD_substr(langs, aoffset, alength)
1.145.4.3 paf 189: :CORD_chars((char)lang, alength);
1.145 paf 190: }
1.145.4.1 paf 191:
1.145.4.3 paf 192: /// appending when 'langs' already contain something [simple cases hanled elsewhere]
1.145.4.4 paf 193: template<typename C>
194: void append(C current,
1.145.4.1 paf 195: const CORD to_nonempty_target_langs) {
1.145.4.3 paf 196: assert(langs);
1.145.4.1 paf 197:
198: if(is_not_just_lang)
1.145.4.9! paf 199: langs=CORD_cat(langs, to_nonempty_target_langs);
1.145.4.1 paf 200: else { // we were "just lang"
1.145.4.4 paf 201: size_t current_size=get_length(current);
1.145.4.1 paf 202: assert(current_size);
1.145.4.9! paf 203: langs=CORD_cat(
1.145.4.3 paf 204: CORD_chars((char)lang, current_size), // first piece [making from just 'lang']
1.145.4.1 paf 205: to_nonempty_target_langs); // new piece
206: }
1.145 paf 207: }
1.145.4.1 paf 208:
1.145 paf 209: public:
1.145.4.1 paf 210:
211: Languages(): langs(0) {}
212: Languages(Language alang): lang(alang), is_not_just_lang(0) {}
213:
1.145.4.4 paf 214: /// MUST be called exactly prior to modification of current [uses it's length]
215: template<typename C>
216: void append(C current, Language alang, size_t asize) {
1.145.4.1 paf 217: assert(alang);
218: assert(asize);
219:
1.145.4.3 paf 220: if(!is_not_just_lang)
221: if(lang) {
222: if(lang==alang) // same length? ignoring
223: return;
224: } else {
225: lang=alang; // to uninitialized
226: return;
227: }
1.145.4.1 paf 228:
1.145.4.4 paf 229: append(current, CORD_chars((char)alang, asize));
1.145 paf 230: }
231:
1.145.4.4 paf 232: /// MUST be called exactly prior to modification of current [uses it's length]
233: template<typename C>
234: void append(C current, size_t appending_length,
1.145.4.1 paf 235: const Languages src) {
1.145.4.4 paf 236: assert(appending_length);
1.145.4.1 paf 237:
1.145.4.3 paf 238: if(!langs)
239: langs=src.langs; // to uninitialized
240: else if(!src.is_not_just_lang)
1.145.4.4 paf 241: append(current, src.lang, appending_length); // simplifying when simple source
1.145.4.3 paf 242: else
1.145.4.4 paf 243: append(current, src.make_langs(appending_length));
1.145.4.1 paf 244: }
245:
1.145.4.4 paf 246: /// MUST be called exactly prior to modification of current [uses it's length]
247: template<typename C>
248: void append(C current,
1.145.4.1 paf 249: const Languages src, size_t aoffset, size_t alength) {
250: assert(alength);
251:
1.145.4.3 paf 252: if(!langs) // to uninitialized?
1.145.4.1 paf 253: if(src.is_not_just_lang)
1.145.4.3 paf 254: langs=CORD_substr(src.langs, aoffset, alength); // to uninitialized complex
255: else
256: lang=src.lang; // to uninitialized simple
257: else
258: if(!is_not_just_lang && !src.is_not_just_lang && lang==src.lang) // both simple & of same language?
259: return; // ignoring
1.145.4.1 paf 260: else
1.145.4.4 paf 261: append(current, src.make_langs(aoffset, alength));
1.145.4.1 paf 262: }
263:
1.145.4.5 paf 264: /// checks if we have lang<=alang all from aoffset to aoffset+alength
1.145.4.1 paf 265: bool check_lang(Language alang, size_t aoffset, size_t alength) const {
266: if(alang==L_UNSPECIFIED) // ignore lang?
267: return true;
268:
269: if(is_not_just_lang)
1.145.4.5 paf 270: return CORD_range_contains_chr_greater_then(langs, aoffset, alength, (unsigned)alang)==0;
1.145.4.1 paf 271: else
1.145.4.5 paf 272: return lang<=alang;
1.145.4.1 paf 273: }
274:
1.145.4.4 paf 275: template<typename C, typename I>
276: void for_each(C current,
1.145.4.2 paf 277: int callback(char, size_t, I), I info) const {
1.145.4.1 paf 278:
279: if(is_not_just_lang)
1.145.4.2 paf 280: CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info);
1.145.4.1 paf 281: else
1.145.4.4 paf 282: callback(lang, get_length(current), info);
283: }
284:
285: bool invariant(size_t current_length) {
286: if(!langs)
287: return current_length==0;
288: if(is_not_just_lang)
1.145.4.7 paf 289: return CORD_len(langs)==current_length;
1.145.4.4 paf 290: return true; // uncheckable, actually
1.145 paf 291: }
292: };
293:
294: struct C {
295: const char *str;
296: size_t length;
297: operator const char *() { return str; }
298: C(const char *astr, size_t asize): str(astr), length(asize) {}
299: };
300:
301: struct Cm {
302: char *str;
303: size_t length;
304: //operator char *() { return str; }
305: Cm(char *astr, size_t asize): str(astr), length(asize) {}
306: };
307:
308: private:
309:
1.145.4.1 paf 310: Languages langs; ///< string characters lang
1.145 paf 311: StringBody body; ///< all characters of string
312:
1.145.4.4 paf 313: #define ASSERT_STRING_INVARIANT(string) \
314: assert((string).langs.invariant((string).body.length()))
315:
1.8 paf 316: public:
317:
1.145 paf 318: explicit String(const char* cstr=0, size_t helper_length=0, bool tainted=false);
319: explicit String(const C cstr, bool tainted=false);
1.145.4.4 paf 320: String(StringBody abody, Language alang): body(abody), langs(alang) {
321: assert(!body.is_empty());
322: ASSERT_STRING_INVARIANT(*this);
323: }
324: String(const String& src): body(src.body), langs(src.langs) {
325: ASSERT_STRING_INVARIANT(*this);
326: }
1.145 paf 327:
328: /// for convinient hash lookup
329: operator const StringBody() const { return body; }
330:
331: bool is_empty() const { return body.is_empty(); }
332: size_t length() const { return body.length(); }
333:
334: /// convert to CORD. if 'lang' known, forcing 'lang' to it
335: StringBody cstr_to_string_body(Language lang=L_AS_IS,
336: SQL_Connection* connection=0,
337: const Request_charsets *charsets=0) const;
338:
339: /// convert to constant C string. if 'lang' known, forcing 'lang' to it
340: const char* cstr(Language lang=L_AS_IS,
341: SQL_Connection* connection=0,
342: const Request_charsets *charsets=0) const {
343: return cstr_to_string_body(lang, connection, charsets).cstr();
344: }
345: /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it
346: char *cstrm(Language lang=L_AS_IS,
347: SQL_Connection* connection=0,
348: const Request_charsets *charsets=0) const {
349: return cstr_to_string_body(lang, connection, charsets).cstrm();
1.50 paf 350: }
1.108 parser 351: /// puts pieces to buf
1.145 paf 352: Cm serialize(size_t prolog_size) const;
1.108 parser 353: /// appends pieces from buf to self
1.145 paf 354: bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
355: /// @see StringBody::append_know_length
356: String& append_know_length(const char* str, size_t known_length, Language lang);
357: /// @see StringBody::append_help_length
358: String& append_help_length(const char* str, size_t helper_length, Language lang);
359: String& append_strdup(const char* str, size_t helper_length, Language lang);
360:
361: bool operator == (const char* y) const { return body==StringBody(y); }
362: bool operator != (const char* y) const { return body!=StringBody(y); }
363:
364: /// this starts with y
365: bool starts_with(const char* y) const {
366: return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(y))==0;
367: }
368: /// x starts with this
369: bool this_starts(const char* x) const {
370: return StringBody(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
1.26 paf 371: }
372:
1.145 paf 373: String& append_to(String& dest, Language lang, bool forced) const;
374: String& append(const String& src, Language lang, bool forced=false) {
375: return src.append_to(*this, lang, forced);
376: }
377: String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
378: String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
379: String& operator << (const StringBody src) {
1.145.4.1 paf 380: langs.append(body, L_AS_IS, src.length());
1.145 paf 381: body<<src;
382: return *this;
383: }
1.100 parser 384:
1.142 paf 385: /// extracts first char of a string, if any
386: char first_char() const {
1.145 paf 387: return is_empty()?0:body.fetch(0);
1.142 paf 388: }
1.54 paf 389:
1.145 paf 390: bool operator < (const String& src) const { return body<src.body; }
391: bool operator > (const String& src) const { return body>src.body; }
392: bool operator <= (const String& src) const { return body<=src.body; }
393: bool operator >= (const String& src) const { return body>=src.body; }
394: bool operator != (const String& src) const { return body!=src.body; }
395: bool operator == (const String& src) const { return body==src.body; }
396:
1.54 paf 397: /// extracts [start, finish) piece of string
1.145 paf 398: String& mid(size_t substr_begin, size_t substr_end) const;
399:
400: /**
401: ignore lang if it's L_UNSPECIFIED
402: but when specified: look for substring that lies in ONE fragment in THAT lang
403: @return position of substr in string, -1 means "not found" [const char* version]
404: */
405: size_t pos(const StringBody substr,
406: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
407: /// String version of @see pos(const char*, int, Language)
408: size_t pos(const String& substr,
409: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
410: size_t pos(char c,
411: size_t this_offset=0) const {
412: return body.pos(c, this_offset);
413: }
1.55 paf 414:
1.145 paf 415: void split(ArrayString& result,
416: size_t& pos_after,
417: const char* delim,
418: Language lang=L_UNSPECIFIED, int limit=-1) const;
419: void split(ArrayString& result,
420: size_t& pos_after,
1.62 paf 421: const String& delim,
1.145 paf 422: Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 423:
1.145 paf 424: typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 425: int prestart, int prefinish,
426: int poststart, int postfinish,
1.68 paf 427: void *info);
1.87 parser 428: /**
1.145 paf 429: @return table of found items, if any.
1.87 parser 430: table format is defined and fixed[can be used by others]:
431: @verbatim
432: prematch/match/postmatch/1/2/3/...
433: @endverbatim
434: */
1.145 paf 435: Table* match(Charset& source_charset,
1.64 paf 436: const String& regexp,
1.145 paf 437: const String* options,
1.99 parser 438: Row_action row_action, void *info,
1.145 paf 439: bool& just_matched) const;
1.87 parser 440: enum Change_case_kind {
441: CC_UPPER,
442: CC_LOWER
443: };
1.145 paf 444: String& change_case(Charset& source_charset,
1.87 parser 445: Change_case_kind kind) const;
1.145 paf 446: const String& replace(const Dictionary& dict) const;
1.96 parser 447: double as_double() const;
448: int as_int() const;
1.137 paf 449:
1.7 paf 450: private: //disabled
451:
1.12 paf 452: String& operator = (const String&) { return *this; }
1.7 paf 453:
1.1 paf 454: };
1.145.4.4 paf 455:
456: template<>
457: inline size_t get_length<StringBody>(StringBody body) {
458: return body.length();
459: }
1.119 paf 460:
1.145 paf 461: /// simple hash code of string. used by Hash
462: inline uint hash_code(const StringBody self) {
463: return self.hash_code();
1.119 paf 464: }
1.1 paf 465:
466: #endif
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to pa_string.h CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / include